Since my older son was diagnosed with an autism spectrum disorder (ASD) at age 3, I read many books on autism. From those books I learnt that the chance that a sibling would also have/develop an ASD was about 5%, compared with the 1% chance for anyone in the population (that is, about 1% of children are officially diagnosed with autism, but I think one can seriously doubt whether that figure is not an underestimation due to under-diagnosis).
I always thought that this 5% figure was odd, since it didn’t correspond at all to my observation at the special-needs-daycare/school of my son or in online parent support groups or in accounts of families affected by ASDs that I read, where many parents report to have several children with an ASD. I noticed just way too many children who also had siblings with an ASD to make that figure of 5% correspond to reality. And now, there’s a study just published in Pediatrics, confirming my observation: if a parent has a child with autism, the chance of a sibling also developing an ASD is almost 20%. That’s what the authors found in a large American sample, and I don’t see any reason why it would be different for other parts of the world.
Not sure how that will change the way we look at autism (if it will make any difference at all), but I find it a striking (but not surprising) figure.
{ 23 comments }
Andrew F. 08.16.11 at 10:38 am
I wonder if there are any studies on autism in adults that would further confirm this relationship? I vaguely recall a paper showing that rates of autism among adults are not significantly different from those among children; perhaps that might have some data on sibling correlations.
ajay 08.16.11 at 11:33 am
Apart from anything else, you’d expect autism to be diagnosed correctly more often in younger siblings of autistic kids than in the general population, simply because the parents would already be aware of what autism looked like from their experience with autistic kid no.1 (or, to be more accurate, diagnosed-autistic kid no. 1).
Red 08.16.11 at 12:34 pm
It is a striking figure that finally confirms what parents confronted with autism already knew or at least suspected. How does it change the way we address autism? I should think broad awareness of the higher risk among health professionals and educators will make early detection, and thus early treatment, more likely. The figure is likely to go up even more, though hopefully with a better outcome for the children.
understudy 08.16.11 at 12:59 pm
Interesting – reminds me a bit of the below study that showed autism rates in Korea at 1:38. At what point do we switch viewing ASD as a disorder and treat it more like we do human sexuality – a spectrum that everyone is on, not a gay/straight, yes/no?
http://opac.yale.edu/news/article.aspx?id=8528
Barry 08.16.11 at 1:13 pm
Seconding ajay – once a family has one member with a frequently-missed condition, they’ll probably spot another member with far higher probability.
Chris Bertram 08.16.11 at 2:17 pm
I think there may be something to understudy’s point here. I suspect that because we are talking about a spectrum, there will be really big differences among different medical cultures about this, with some being willing to diagnose quite a high proportion as suffering from a disorder and others placing persons with the same dispositions and behaviours in the “normal” population. So perhaps quite hard to get any definite fact of the matter.
ajay 08.16.11 at 2:53 pm
As for this question:
Not sure how that will change the way we look at autism (if it will make any difference at all), but I find it a striking (but not surprising) figure.
Probably “not very much”, to be honest. It suggests that there’s a much higher heritable component in autism than we thought.
Or maybe not; even if autism’s entirely environmental then it would also make sense that two kids brought up in the same house, eating the same food and so on would both get it. I’m sure you’re more likely to break your collarbone falling off a horse if your older sibling has done so, but that’s just another way of saying that some families are horsey and some families aren’t.
But will this necessarily change the way we regard autism? Not as much as the gradual spread of the idea of an “autistic spectrum”, as CB hints.
The next step might be extending the autistic spectrum beyond normal on the other side; maybe it’s possible to be ‘too unautistic’ – too unfocussed, for example.
Ed Marshall 08.16.11 at 3:53 pm
My son was diagnosed with Pervasive Developmental Disorder and it scared me so badly that I got myself vasectomized. Later the diagnosis slipped around to Asperger’s. By the time he was six, he just seemed like an extremely intelligent very witty kid. As a matter of fact, at 16 he looks, talks and acts almost exactly like me. He seems to value the Asperger’s designation and likes to talk on a forum dedicated to Asperger’s disorder members, but I think part of the growth of autism is a weakening definition. I have no doubt that if I was born 20 years later than I was that I would have been given this designation.
Chris Johnson 08.16.11 at 5:16 pm
As ajay points out, heritability vs environmental influences can bedevil observational studies such as this one. Twin studies are one way to address that issue, and this has been done. Although I am a pediatrician, I don’t follow this literature closely. But at least one study compared autism rates in identical vs fraternal twins. The concordance rate between twins for a diagnosis for autism was much, much higher among the identical twins. Using the looser criterion of social and cognitive difficulties, there was a 92% concordance in the identical twins, 10 % among the fraternal. The authors concluded that genetics played the dominant role. Here’s a link to the abstract for those interested. The study is 15 years old, and is reference #3 in the Pediatrics article linked in the OP.
http://www.ncbi.nlm.nih.gov/pubmed/7792363
bianca steele 08.16.11 at 6:45 pm
I filled out an autism-diagnostic questionnaire at my daughter’s 24-month and either 12- or 18-month appointments, within the past couple of years. (I think this is the questionnaire that was used in a recent study, though not the one the OP links.) The earlier questionnaire had two or three times more questions on it, but there was not excessive overlap between either set of questions and the markers for autism I would have picked up from reading general-audience books and encyclopedia articles on the subject. The former were also heavy on anecdotes, which seemed to be of questionable informational value.
vivian 08.16.11 at 8:05 pm
As a consequence of speaking of a continuum, parents diagnose ourselves more severely than people who know us do. Though most of us would have benefitted from some special services, that is probably true for everyone.
Insofar as things like pragmatic speech therapy help raise the floor for all kids, I don’t care much whether it is 1:160 or 1:40 or 1:2. Whether having more kids ‘on the spectrum’ makes the education budget pie bigger or the slices thinner I do not know. Probably varies.
Sock Puppet of the Great Satan 08.16.11 at 8:19 pm
“It suggests that there’s a much higher heritable component in autism than we thought.”
Also, there’s an incentive for parents to Get On That Problem ASAP with later children, given that early intervention with autistic kids yields better eventual outcomes.
As wanting as therapies for autistic kids are now, one feels pain for what earlier generations of autistic kids lumped in various “mentally defective” categories went through.
“The concordance rate between twins for a diagnosis for autism was much, much higher among the identical twins. ”
That still doesn’t rule out environmental factors entirely. If the cause is due to some epigenetic interaction of some environmental factor (say, certain anticonvulsants: see http://www.sciencedirect.com/science/article/pii/S0890856710003916) and a genetic susceptibility causing that gene (say, the oxytocin receptor gene) to be more susceptible to methylation, you still may have an environmental component which wouldn’t necessarily show up in fraternal twin/identical twin studies because of the similarity of the environment the identical twins were exposed to. Which would explain an increasing rate of diagnosis even if the genetic background were unchanged.
Red 08.16.11 at 9:27 pm
I object to the notion that this is just a matter of vague borders, a question of “judgment”. Sure, some individuals with Asperger’s may be able to “function” (hate that word) just fine in our world. But that number is relatively small. And autism is not comparable to ASD.
Of course both environmental and inherited factors play a role.
Steve LaBonne 08.16.11 at 10:17 pm
And I object to the notion that I’m not rich. Sadly, in both cases the thing being objected to happens to be the truth.
Chris Johnson 08.16.11 at 10:41 pm
“Of course both environmental and inherited factors play a role.”
Except for some clearly genetic disorders, this is the case for most diseases I can think of, including serious infections.
An important part of the context for studies like this the one in Pediatrics is the recent Wakefield MMR vaccine hoax, because that unfortunate incident made much of the public think outside, environmental factors were the cause of autism. So this study is useful. As the OP points out, parents of autistic children already suspected a strong genetic component.
Chrisb 08.16.11 at 11:30 pm
The problem with autism is that we don’t have a clear description of what we’re talking about. The DSM-IV takes a chinese menu approach to the condition, saying one from column A and three from column B in a fashion that I’ve calculated means that (working out the permutations and combinations) there are over 12,000 different ways to be autistic, before you even start looking at level of severity. We know what type cases look like – Rain Man was enormously influential – but it’s a lot harder to give a set of criteria that catch people with autism and only people with autism – present diagnostic criteria are designed to let nobody escape who might be autistic, not to rule out people who might be normal, and they’re bound to overpredict; the variable is not so much the number who meet criteria but the number who are brought to the diagnostician. Which means, too, that to say that “early intervention with autistic kids yields better eventual outcomes”, even if it was true, would not settle the issue; you’d also have to weigh up the cases of early intervention with kids who turned out to be misdiagnosed (if, that is, these weren’t the children whose improvement was already being counted into those ‘better outcomes’, which is also likely).
Bear in mind, too, that a good chunk of the rise in diagnosed autism can be accounted for by a transfer across from the population that would formerly have been diagnosed with simple intellectual disability. And this has to mean that Red’s attempt to say that all autism and most Asbergers involve an inability to function just can’t stand. The number of people with those diagnoses who are that severely affected – little or no speech, little or no social interaction — is still very small (not, for example, applying to one Korean in 38). The penumbra grows a lot faster than the core.
When we know more about this, if we ever do, we’ll be talking about the autisms. In the meantime we should at least acknowledge that our grasp of the concept is fuzzy and amorphous, and that any clear statement is ipso facto improbable.
liberal 08.17.11 at 12:37 am
ajay wrote, “Or maybe not; even if autism’s entirely environmental then it would also make sense that two kids brought up in the same house, eating the same food and so on would both get it. ”
Nope. It’s very likely that autism develops prenatally. Of course, that still leaves lots of room for environmental influences.
Jonathan H 08.17.11 at 3:56 am
There was a fascinating study done in the last year that showed how autism diagnoses were rising rapidly in higher income households, but total mental disability diagnoses were not rising. In other words, autism diagnoses were very likely being given instead of diagnoses that had a more pejorative connotation (mental retardation), but this was occurring in upper income brackets while historical patterns of diagnosis were largely unchanged for lower income households.
I think the study may have been done in Australia, but I couldn’t find it on a quick search. I agree with the several commenters who pointed out that autism is a vague collection of symptoms and not a clear biological condition, and there is evidence for, if not overdiagnosis, at least culturally-sensitive diagnosis.
herr doktor bimler 08.17.11 at 6:25 am
If the cause is due to some epigenetic interaction of some environmental factor (say, certain anticonvulsants: see http://www.sciencedirect.com/science/article/pii/S0890856710003916) and a genetic susceptibility causing that gene (say, the oxytocin receptor gene) to be more susceptible to methylation, you still may have an environmental component which wouldn’t necessarily show up in fraternal twin/identical twin studies because of the similarity of the environment the identical twins were exposed to.
If I understand correctly, you are proposing that there are alleles for particular genes that make developing foetuses or infants who carry the allele more vulnerable to an environmental factor. In this scenario, identical twins show a higher concordance than fraternal twins for the impact of exposure to the factor. For the purposes of the scenario, it does not matter whether the vulnerability has an epigenetic mechanism (i.e. through methylation of the gene promoter area after its initial demethylated state).
Aren’t you just saying that (a) genome and environment interact so causation is not simply one-or-the-other, and (b) identical / fraternal twin comparisons can identify genetic causes but don’t tell us anything about environmental causes?
Come to think of it, these are both probably worth re-stating.
I skimmed the J. Am. Ac. Ch & Adol. Psychiat. paper but it seemed to have a high ratio of speculation to fact. The authors list a number of genetic disorders where maternal imprinting is involved, then elide this with the cellular-differentiation aspect of epigenetics as if the two are the same. The paper seemed to arise from the journal’s decision to have a special issue about epigenetics rather than from any empirical need.
Sock Puppet of the Great Satan 08.18.11 at 6:25 pm
herr doktor wrote:
“Aren’t you just saying that (a) genome and environment interact so causation is not simply one-or-the-other, and (b) identical / fraternal twin comparisons can identify genetic causes but don’t tell us anything about environmental causes?”
That’s about it. IANAG (I am not a geneticist), but if the cause is binding of one or more (genetically determined) receptor(s) A to an (environmental) agent B, because of the similarity of the interuterine and post-birth environments, a fraternal/identical twin comparison study is going to pick up the absence or presence of A but not the absence or present of B. So the effect of the environment would be underestimated. It’s just the purely genetic explanation doesn’t explain the increasing rates of autism, and shifts in diagnosis (in my view) don’t quite explain the increase entirely away. There was a hypothesis that consumption of folic acid during pregnancy was a potential cause a few years back: I don’t know whether that’s been disproven, but it had circumstantial evidence in favor.
Tried to find a good recent review article on epigenetics of autism, but this 2006 article was the best I could do:
http://www.sciencedirect.com/science/article/pii/S0166223606001111
I welcome the focus on the genetics of autism after the medical and scientific community have had to fight two decades of ad-hoc hypotheses linking vaccination and (it’s thimerosol! no it’s MMR! no, it’s giving lots of vaccinations at once! etc.), though.
PaulB 08.18.11 at 8:55 pm
I have no view on what the true probability is of parents with one autistic child having another autistic child. But I am sure that parents reporting online do not constitute a representative sample.
herr doktor bimler 08.18.11 at 11:20 pm
Sock Puppet of the Great Satan said
Thanks. I think that journal is on my library’s electronic-access list.
I’ve become cautious about the word “epigenetics” because some of the anti-vax loons have glommed onto it as a way of saving their favoured ‘thimerosol’ theory. The claim used to be that a mercury-based antibiotic in some vaccines was responsible for autism… so the antibiotic was removed… when autism rates failed to fall as predicted, the campaigners simply responded that the problem was with the vaccines administered to the mothers — the mercury having altered their DNA in some way that afflicted the next generation (oh yes, and current vaccines are still evil for some reason yet to be determined). Hence ‘epigenetics’.
You and I know that this isn’t how epigenetics work, but the anti-vaccine crowd have their own cargo-cult understanding of science and are primarily interested in the brightness and shininess of the word.
Anon 08.20.11 at 1:05 am
I lived independently all my life and never got diagnosed with Asperger’s syndrome or autism, though I have shown signs and experienced difficulties stereotypical to it all along. Now children with early development histories similar to mine are being labeled autistic. This might benefit some of them, but it is also exposing them to quackery and putting their parents through needless anxiety. And it’s probably causing far too much money to be diverted to them when it is better spent helping those families whose children will not be able to enter legal adulthood on schedule.
Comments on this entry are closed.